English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Adversarial example detection is known to be an effective adversarial defensemethod. Black-box attack, which is a more realistic threat and has led tovarious black-box adversarial training-based defense methods, however, does notattract considerable attention in adversarial example detection. In this paper,we fill this gap by positioning the problem of black-box adversarial exampledetection (BAD). Data analysis under the introduced BAD settings demonstrates(1) the incapability of existing detectors in addressing the black-box scenarioand (2) the potential of exploring BAD solutions from a data perspective. Totackle the BAD problem, we propose a data reconstruction-based adversarialexample detection method. Specifically, we use variational auto-encoder (VAE)to capture both pixel and frequency representations of normal examples. Then weuse reconstruction error to detect adversarial examples. Compared with existingdetection methods, the proposed method achieves substantially better detectionperformance in BAD, which helps promote the deployment of adversarial exampledetection-based defense solutions in real-world models.

Towards Black-box Adversarial Example Detection: A Data Reconstruction-based Method